Forward Vehicle Detection Using ROI Search and Feature Tracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle detection systems face challenges in accurately and stably detecting forward vehicles from captured driving images, which is crucial for enhancing the accuracy of advanced driver assistance systems (ADAS) like forward vehicle collision warning systems (FVCWS).
Innovation Solution
A method involving continuous reception of front images, setting a search area for vehicle detection based on the location or area detected in previous images, using a machine learning model for vehicle detection, and tracking vehicles using feature points extracted from previous images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If vehicle detection is performed using machine learning model on entire captured images, then detection accuracy is improved, but processing speed deteriorates
Solution Approach 1:
The captured image is divided into multiple regions of interest (ROIs) based on predicted vehicle locations from previous frames. The machine learning model is applied only to these segmented ROIs rather than the entire image, reducing computational load while maintaining detection accuracy in critical areas.
Solution Approach 2:
Vehicle prediction based on previous frame detection results is performed before actual detection. This preliminary action identifies likely vehicle locations, allowing the system to focus processing resources on these areas and avoid processing entire images, thus improving processing speed without sacrificing accuracy.
2Reliability
If search area is enlarged to ensure vehicle detection, then detection reliability is improved, but processing time increases
Solution Approach 1:
Different search area sizes are applied to different regions based on predicted vehicle likelihood. High-confidence prediction regions receive focused search areas, while low-confidence regions receive expanded search areas. This local differentiation maintains reliability where needed while reducing overall processing time.
Solution Approach 2:
The search area is dynamically adjusted to cover only the necessary portion of the image based on prediction confidence. Instead of always processing the entire image or fixed large search areas, the system applies partial processing to regions where vehicles are likely to be found, reducing detection time while maintaining reliability through targeted search.
Data Source
AI summary
There is provided a method for detecting a vehicle including receiving continuously captured front images, setting a search area of the vehicle in a target image based on a location of the vehicle or a vehicle area detected from a previous image among the front images, detecting the vehicle in the search area according to a machine learning model, and tracking the vehicle in the target image by using feature points of the vehicle extracted from the previous image according to a vehicle detection result based on the machine learning model. Since the entire image is not used as a vehicle detection area, a processing speed may be increased, and a forward vehicle tracked in an augmented reality navigation may be continuously displayed without interruption, thereby providing a stable service to the user.